Thermal degradation of poly(L-lactide): effectof alkali earth metal oxides for selectiveL,L-lactide formation

著者 Fan Yujiang, Nishida Haruo, Mori Tomokazu,Shirai Yoshihito, Endo Takeshi

journal orpublication title

Polymer

volume 45number 4page range 1197-1205year 2004-01-10URL http://hdl.handle.net/10228/00006758

doi: info:doi/10.1016/j.polymer.2003.12.058

1

Thermal Degradation of Poly (L-lactide): Effect of Alkali Earth Metal Oxides for

Selective L,L-Lactide Formation

Yujiang Fana,b, Haruo Nishidaa,*, Tomokazu Moria,c, Yoshihito Shiraib, and Takeshi

Endoa,d

a Molecular Engineering Institute, Kinki University, 11-6 Kayanomori, Iizuka, Fukuoka

820-8555, Japan

b Graduate School of Life Science and Systems Engineering, Kyushu Institute of

Technology, 1-1 Hibikino, Kitakyushu, Fukuoka 808-0196, Japan

c Faculty of Computer Science and Systems Engineering, Kyushu Institute of

Technology, 680-4 Kawazu, Iizuka, Fukuoka, 820-8502, Japan

d Faculty of Engineering, Yamagata University, 4-3-16 Jonan, Yonezawa, Yamagata

992-8510, Japan

*Corresponding author: Haruo Nishida

Molecular Engineering Institute, Kinki University, 11-6 Kayanomori, Iizuka,

Fukuoka 820-8555, Japan

Tel/Fax: +81-948-22-5706.

E-mail address: hnishida@mol-eng.fuk.kindai.ac.jp (H. Nishida)

© 2004. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/

2

Abstract

To achieve the feed stock recycling of poly(L-lactide) (PLLA) to L,L-lactide,

PLLA composites including alkali earth metal oxides, such as calcium oxide (CaO) and

magnesium oxide (MgO), were prepared and the effect of such metal oxides on the

thermal degradation was investigated from the viewpoint of selective L,L-lactide

formation. Metal oxides both lowered the degradation temperature range of PLLA and

completely suppressed the production of oligomers other than lactides. CaO markedly

lowered the degradation temperature, but caused some racemization of lactide,

especially in a temperature range lower than 250 °C. Interestingly, with MgO

racemization was avoided even in the lower temperature range. It is considered that the

effect of MgO on the racemization is due to the lower basicity of Mg compared to Ca.

At temperatures lower than 270 °C, the pyrolysis of PLLA/MgO (5 wt %) composite

occurred smoothly causing unzipping depolymerization, resulting in selective L.L-

lactide production. A degradation mechanism was discussed based on the results of

kinetic analysis. A practical approach for the selective production of L,L-lactide from

PLLA is proposed by using the PLLA/MgO composite.

Keywords: Thermal degradation / Poly(L-lactide) / Racemization

3

1. Introduction

Poly(L-lactide), or poly(L-lactic acid) (PLLA) belongs to a group of

biodegradable polymers and has received much interest because of its medical,

pharmaceutical, and environmental applications [1-3]. Nowadays, PLLA and its related

copolymers are attracting much attention as promising alternatives to the normal

petroleum based commodity resins, because they can be derived from renewable

resources, such as corn, potato and other agricultural products and have many useful

properties, such as mechanical strength, transparency, and compostability [4,5].

However, whilst the biodegradability of PLLA is excellent in terms of its ability to be

bioabsorbed, its microbial degradation is limited to a few species of microorganisms,

such as Amycolatopsis and Streptomyces strains [6,7]. Any large-scale consumption of

PLLA products will bring the associated problem of an excess of PLLA waste, which

will be difficult to treat by biodegradation either in composting plants or in the natural

environment. One possible approach to overcoming this difficulty is to regenerate the

cyclic monomer, L,L-lactide, from the PLLA waste, using the well-known fact that

PLLA can be changed into the monomer by thermal degradation [8-15]. However, the

thermal degradation of PLLA has been reported to be more complex than the simple

reaction that gives rise to lactide [11-15]. McNeill and Leiper investigated the

degradation of PLLA under both controlled heating conditions and isothermal

conditions [8,9]. They reported that the main products were cyclic oligomers, including

lactide. Other lower boiling point products, such as carbon dioxide, acetaldehyde,

ketene, and carbon monoxide were also produced. Kopinke and co-workers reported a

multi-step process for PLLA pyrolysis. They found that the intra-molecular

transesterification was a dominant degradation pathway, and that pyrolysis behavior

4

was different between pure as opposed to Sn-containing PLLAs [11,12]. In addition,

considerable racemization to meso- and D,D-lactides was observed during the pyrolysis

of PLLA [11,12,16,17], which caused serious problems after the reproduction of PLLA,

diminishing its crystallizability and some other useful properties [18-20]. Racemization

during pyrolysis has scarcely been discussed in previously published works, except that

of Kopinke et al., who proposed a racemization mechanism through an ester-semiacetal

tautomerization occurring in a lactate unit in the chain under higher temperatures [12].

This speculation is reasonable as it explains the fact that more than two

diastereoisomers were observed for each cyclic oligomer as pyrolyzates. Generally,

temperature and catalyst are regarded as affecting the pyrolysis significantly. However,

no detailed and quantitative discussion on these points has taken place so far. Thus, it is

necessary to clarify the racemization mechanism in the pyrolysis, and find an

appropriate approach for controlling the degradation reaction of PLLA so as to

reproduce the optically pure L,L-lactide.

It has been reported that the blending of many kinds of metal compounds with

PLLA greatly influences the pyrolysis behaviour [21,22]. Noda et al. evaluated the

activity of a series of Al, Ti, Zn, Zr, and Sn compounds as intramolecular

transesterification catalysts for the pyrolysis of PLLA oligomer, and reported that the

activity of each metal was in the following order: Sn>Zn>Zr>Ti>Al [21]. Cam et al.

also reported that compounds of metals, such as Sn, Zn, Al, and Fe, had a great

influence on the pyrolysis behavior of PLLA [22]. Though the Sn compounds are

effective catalysts for L.L-lactide production from PLLA [23], less toxic and more easily

available catalysts should be used for feedstock recycling.

5

In our previous studies [24.25], it was found that the calcium salt end capped

PLLA, PLLA-Ca, degraded through unzipping depolymerization to produce principally

lactides. Alkali earth metal compounds are ideal catalysts for recycling, because of their

lower toxicity and easy availability. Kinetic analysis of the pyrolysis process revealed

that there was one selective L,L-lactide formation and two racemization reactions

occurring in different temperature ranges each with its own specific mechanism [25]. At

temperatures lower than 250 °C, a SN2 attack from carboxylate anion on an

asymmetrical carbon dominates, and at higher temperatures than 320 °C ester-

semiacetal tautomerization occurs to produce meso- and D,D-lactides. In a temperature

range of 250-320 °C, the unzipping depolymerization mechanism dominates, resulting

in selective L,L-lactide formation. These temperature-dependent multiple mechanisms

have been revealed as occurring in the pyrolysis of the homogeneous PLLA-Ca. If these

multiple mechanisms are applied to heterogeneous composites of PLLA and inorganic

compounds, then this approach will be usable for the practical feedstock recycling of

PLLA.

In this paper, two alkali earth metal oxides, calcium oxide (CaO) and magnesium

oxide (MgO), are blended with PLLA to prepare the PLLA/metal oxide composites.

These metal oxides are naturally existing materials, and in particular MgO is commonly

used in large quantities as a filler for plastics. The particle surfaces of these metal oxides

are expected to have catalytic effects on the PLLA pyrolysis. The thermal degradation

behavior of the composites is examined as a function of temperature. Consequently, the

metal oxides in the composites markedly influenced the degradation temperature,

kinetics, and racemization of the PLLA pyrolysis. A practical approach usable in

controlling racemization and obtaining optically pure L,L-lactide is demonstrated.

6

2. Experimental

2.1 Materials

Monomer, L,L-lactide, was obtained from Shimadzu Co. Ltd. and purified by

being recrystallized three times from dry toluene and one time from dry ethyl acetate.

The vacuum dried L,L-lactide was stored in N2 atmosphere. After the purification, no

meso-lactide was detected by gas chromatography (GC). Polymerization catalyst, Sn(II)

2-ethylhexanoate, Sn(Oct)2, obtained from Wako Pure Chemical Industries, Ltd. was

distilled under reduced pressure before use. Calcium oxide (CaO, average diameter H 10

µm, 99.0 %) and magnesium oxide (MgO, average diameter H 0.01µm, 99.9 %) were

purchased from Wako Pure Chemical Industries, Ltd. and used as received. Other

chemicals and solvents are used without further purification.

2.2 Measurements

Molecular weight was measured by gel permeation chromatography (GPC) on a

TOSOH HLC-8220 GPC system at 40 °C using TOSOH TSKgel Super HM-H column

and a chloroform eluent (0.6mL min-1). Low polydispersity polystyrene standards with

Mn from 5.0×102 to 1.11×106 were used for calibration.

The residual metal content in the PLLA samples was measured with a Shimadzu

AA-6500F atomic absorption flame emission spectrophotometer (AA). The samples

were degraded by a 25 % ammonia solution, dissolved in 1M-hydrochloric acid, and

then measured by AA.

Thermogravimetric analysis (TG) was conducted on a Seiko Instrumental Inc.

EXSTAR 6200 TG/DTA system in an aluminum pan under a constant nitrogen flow

7

(100 mL min-1) using about 5 mg of the PLLA film sample. Two heating program

patterns were applied for pyrolysis of PLLA, which were, (1) dynamic heating and (2)

isothermal heating. In the dynamic heating process, for each sample, prescribed heating

rates φ of 1, 3, 5, 7, and 9 K min-1 were applied from room temperature to 400 °C. In the

isothermal heating process, the sample was heated from 60 to 250 ºC at a heating rate of

20 ºC min-1 and kept at 250 °C for 10 min. The pyrolysis data were collected at regular

intervals (about 20 times K-1) by an EXSTAR 6000 data platform, and recorded into an

analytical computer system.

Gas chromatography (GC) measurements were recorded on a Shimadzu GC-9A

gas chromatograph with a Varian cyclodextrine-² -236M-19 capillary column (30m x

0.25mm i.d.; film thickness, 0.25µm) at 150 °C, using helium as the carrier gas. The

peaks for meso-, L,L-, and D,D-lactide in the GC chromatogram were identified by

comparison with the peaks for pure substances.

Pyrolysis-gas chromatograph/mass spectra (Py-GC/MS) were measured on a

Frontier Lab PY-2020D double-shot pyrolyzer connected to a Shimadzu GCMS-

QP5050 chromatograph/mass spectrometer, which was equipped with an Ultra Alloy+-5

capillary column (30m x 0.25mm i.d.; film thickness, 0.25µm). High purity helium at

100 mL min-1 was used as a carrier gas. A PLLA sample was put in the pyrolyzer and

heated according to two heating program patterns, (1) dynamic heating and (2)

isothermal heating. In the dynamic heating process, the PLLA sample in the pyrolyzer

was heated from 60 to 400 ºC at a heating rate of 10 ºC min-1. In the isothermal heating

process, the PLLA sample was heated from 60 to 250 ºC at a heating rate of 20 ºC min-1

and kept at 250 °C for 10 min. The volatile pyrolysis products were conducted into the

GC through the selective sampler. The temperature of the column oven was first set at

8

40 °C. After the pyrolysis process had finished, the column was heated according to the

following program: 40 ºC for 1 min; 40-120 ºC at 5 ºC min-1; 120-320 ºC at 20 ºC min-1;

320 °C for 13 min. Mass spectrum measurements were recorded 2 times s-1 during this

period.

2.3 Preparation of PLLA/alkali earth metal oxide composites

PLLA was synthesized by the ring-opening polymerization of L,L-lactide

catalyzed by Sn(Oct)2 as described in the previous reports [24,25]. The obtained raw

PLLA was purified by firstly extracting the catalyst and its residues from the

PLLA/chloroform solution with a 1M HCl aqueous solution, then washing with distilled

water until the aqueous phase became totally neutral, and finally precipitating the

polymer with methanol before vacuum drying (Mn 151,000, Mw/Mn 1.82, Sn content by

AA analysis, 14 ppm - the order of the lower limit of detection under these experimental

conditions). The purified PLLA was dissolved in chloroform and mixed with CaO or

MgO (5 wt % as Ca or Mg to PLLA), and the mixture was vigorously stirred for 1 h to

disperse the inorganic particles uniformly. The dispersed mixture was cast on a glass

Petri dish. The obtained composite film was washed with methanol and then dried in

vacuo for 48 h. Change in the molecular weight of PLLA was scarcely detected after the

preparation (PLLA/CaO: Mn 158,000, Mw/Mn 1.82; PLLA/MgO: Mn 158,000, Mw/Mn

1.87)

3. Results and discussion

3.1 Dynamic pyrolysis of PLLA/CaO and PLLA/MgO composites

9

To examine effects of alkali earth metal oxides as catalysts for PLLA pyrolysis,

CaO and MgO powder was blended with purified PLLA so as to contain 5 wt % of

metal to PLLA, and thermal degradation of the composites was then conducted

gravimetrically on TG/DTA in a N2 flow. Figure 1 shows the TG curves and differential

thermogravimetric (DTG) profiles of PLLA/CaO (5 wt %) and /MgO (5 wt %)

composites, which were measured at heating rate φ = 5 K min-1, and the results then

compared with that of purified PLLA. The weight loss of purified PLLA started at about

270 °C and then proceeded smoothly to reach almost complete degradation at about

370 °C, reproducing the previous results [24]. On the other hand, the TG curves of

composites shifted to rather lower temperature ranges. The TG curve of PLLA/CaO (5

wt %) showed the beginnings of degradation at about 180 °C and a smooth weight loss

to 0.05 in the residual weight ratio, w, at about 260 °C. The weight loss of PLLA/MgO

(5 wt %) began at about 210 °C and reached complete degradation at less than 300 °C,

showing a rapid weight loss curve.

In DTG profiles, purified PLLA showed a main peak found at 357 °C in a higher

temperature range of 320-375 °C and a shoulder at about 300 °C. PLLA/CaO (5 wt %)

also exhibited a wide DTG profile in a temperature range of 180-300 °C, but had a

sharper main peak at 242 °C, suggesting a rapid main degradation process after a

relatively slow initial process. Interestingly, PLLA/MgO (5 wt %) showed a pointed

DTG profile peaking at 280 °C, which indicates a single and very rapid decomposition

process. These results suggest that different decomposition pathways might be taken

during pyrolysis in each case for purified PLLA, PLLA/CaO (5 wt %), and PLLA/MgO

(5 wt %) composites.

[Figure 1 goes here]

10

3.2 Py-GC/MS analysis of pyrolyzates

Pyrolysis products of PLLA/CaO (5 wt %) and /MgO (5 wt %) composites in the

temperature range of 60-400 °C were analyzed by Py-GC/MS. The results are illustrated

in Figure 2. For both the PLLA/CaO (5 wt %) and /MgO (5 wt %) composites, lactides

were detected at 11.8 (meso-lactide) and 13.4 min (L,L- and D,D-lactide) as being the

dominant pyrolysis products, and a small amount of cyclic oligomers other than lactide

was detected in a range of 20 to 30 min. Quantitative summation of the peak intensities

in the Py-GC/MS chromatogram showed that for both the composites the lactides

comprised about 98 % of the degradation products. In contrast, the pyrolysis of purified

PLLA resulted in the production of a large amount of cyclic oligomers through a

random degradation process as reported previously [24]. Thus, the selective lactide

formation on the pyrolysis of the composites suggests that an unzipping

depolymerization is the principal degradation route of PLLA/CaO (5 wt %) and /MgO

(5 wt %) pyrolysis. In the previous report [25], it was assumed that the main pyrolysis

process of a homogeneous material PLLA-Ca with a calcium salt end structure was the

unzipping depolymerization. This proceeded through a back-biting attack from alkoxide

ends upon carbonyl carbons in penultimate lactic units to produce lactide selectively.

Similar reaction pathways may be the main pyrolysis process for the PLLA/CaO (5

wt %) and /MgO (5 wt %) composites.

[Figure 2 goes here]

11

3.3 Racemization of PLLA in pyrolysis of PLLA/CaO and PLLA/MgO composites

In Figure 2, both pyrolyzates from PLLA/CaO (5 wt %) and /MgO (5 wt %)

composites included meso-lactide, which is a cyclic dimer comprised of L- and D-lactate

units. From quantitative calculations of the peak intensity in Figure 2, the pyrolyzates

were found to contribute 17.0 and 7.0 % of meso-lactide to the total lactide for

PLLA/CaO (5 wt %) and /MgO (5 wt %) composites, respectively. This meso-lactide

formation is obviously due to racemization during the thermal degradation, because the

original PLLA was composed of 100 % L-lactate unit.

[Figure 3 goes here]

As reported previously [24,25], the racemization behavior of PLLA-Ca varies

markedly with temperature. To determine the racemization behavior in the pyrolysis of

PLLA/CaO (5 wt %) and /MgO (5 wt %) composites, each pyrolyzate in different

temperature ranges was collected and analyzed with Py-GC/MS. Results of meso-lactide

formation in different temperature ranges are illustrated in Figure 3, in which the meso-

lactide content was calculated from the peak intensities at 11.8 and 13.4 min in Py-

GC/MS chromatograms and the total weight loss derived from TG curves within the

same temperature range. The pyrolysis of PLLA/CaO (5 wt %) showed a rather high

ratio (10-20 %) of racemization at temperatures lower than 250 °C, and a level of

racemization lower than 2 % in the temperature range 250-300 °C. In other words, in

this temperature range L,L-lactide was formed selectively. Though an indication of the

increase was found at 300-320 °C, no significant formation of lactides was observed at

temperatures higher than 300 °C, because the degradation of PLLA/CaO (5 wt %)

finished at about 290 °C.

12

On the other hand, the pyrolysis of PLLA/MgO (5 wt %) composite exhibited a

totally different racemization behavior compared to PLLA/CaO (5 wt %). At

temperatures under 220 °C, volatilized pyrolyzates were scarcely observed. The lactide

formation started at temperatures over 220 °C and the ratio of meso-lactide to the total

lactide in the pyrolyzates was kept at a low level of less than 3 % until 270 °C, showing

selective L,L-lactide formation. Then, at temperatures over 270 °C, the meso-lactide

ratio rose up to about 10 % until the degradation was completed at about 300 °C.

The total meso-lactide content (7 %) of PLLA/MgO (5 wt %) pyrolyzates was less

than 18 % of PLLA/CaO (5 wt %) pyrolyzates. This is considered to be responsible for

the disappearance of the intensive meso-lactide formation at temperatures lower than

250 °C, which was apparent in PLLA/CaO (5 wt %) pyrolysis. The influence of

temperature on the pyrolysis was examined in detail from a viewpoint of kinetics in

following sections.

3.4 Kinetics of PLLA/CaO and PLLA/MgO composites pyrolysis: Activation energy, Ea

To clarify the thermal degradation pathways of PLLA/CaO (5 wt %) and /MgO (5

wt %) composites, a dynamic degradation method was carried out at different heating

rates φ of 1-9 K min-1 in TG/DTA under N2 flow.

At a certain fractional weight ratio w, the apparent activation energy Ea was

determined from the slope of the log(φ) vs 1/T plot according to the following equation

[26,27],

R

bETd

d a−=)/1(

logφ

13

where Ea and R are the apparent activation energy of the thermal degradation and the

molar gas constant, respectively. And, b is a constant in the Doyle’s approximation

equation [28]. Using this approach, results of Ea values for TG data of purified PLLA,

PLLA/CaO (5 wt %), and PLLA/MgO (5 wt %) composites are plotted against w in

Figure 4. It should be noted that these Ea values are apparent values, because not only a

variety of reactions but also a number of physical changes occur during the pyrolysis.

Despite the values being apparent, the Ea value and its changes are effective in

predicting the reaction, because the Ea value most likely reflects the main reaction

occurring at each time and temperature.

[Figure 4 goes here]

The Ea values of purified PLLA steadily increased from 140 kJ mol-1 with a

decrease in w to converge at about 180 kJ mol-1, which is in accordance with previous

results [24]. For PLLA/CaO (5 wt %), during the first half of the pyrolysis the Ea value

increased from 147 to 160 kJ mol-1 with decrease in w, and then decreased to a final

value of 125 kJ mol-1, which is a value relatively near to that for the pyrolysis of PLLA-

Ca (98-120 kJ mol-1) [24]. This indicates that PLLA/CaO (5 wt %) composite changed

its own physical and/or chemical conditions during the pyrolysis, finally approaching a

similar structure to that of PLLA-Ca. On the other hand, the Ea value of PLLA/MgO (5

wt %) was kept at almost a constant value of 120-130 kJ mol-1 during the pyrolysis.

Thus, the thermal degradation of PLLA/MgO (5 wt %) may proceed through a simple

process.

[Figure 5 goes here]

In Figure 5, the Ea values and the total of meso-lactide and cyclic oligomer

contents were plotted against temperature. The three Ea curves for the PLLA/CaO (5

14

wt %), PLLA/CaO (1wt %), and PLLA-Ca (Ca content: 210 ppm) pyrolysis look like

one continuous curve changing with temperature (Figure 5a). Further, this total

combined curve just relates to the meso-lactide content of pyrolyzates in Figure 5b.

These results demonstrate that the pyrolysis process of PLLA/CaO composite comprises

at least three temperature dependent degradation stages, which are the frequent meso-

lactide formation stage at temperatures lower than 250 °C, the dominant L,L-lactide

formation stage in the temperature range of 250-320 °C, and the intensive racemization

and oligomer formation stage at temperatures greater than 320 °C. Almost the same

discussion was made about the pyrolysis process of PLLA-Ca alone in the previous

report [25]. These results indicate that the degradation temperature shifts to a lower

range because of the CaO content, whilst the degradation mechanism remains dependent

on the temperature.

A similar relationship between the Ea values and meso-lactide and cyclic oligomer

content is demonstrated on the pyrolysis of PLLA/MgO (5 wt %) and purified PLLA

(Figures 5c and 5d). In the case of PLLA/MgO (5 wt %) composite, it should be noted

again that a few meso-lactide products are found at temperatures lower than 270 °C,

however, both the Ea value and meso-lactide content gradually increased at temperatures

over 280 °C and look like being linked to those of purified PLLA. The lower

temperature shifts in the Ea value and content may reflect differences in the chemical

properties between MgO and CaO. The production of cyclic oligomers was found only

on the pyrolysis of purified PLLA, because the volatilization of cyclic oligomers is

mainly dependent on the temperature.

15

3.5 Kinetics of PLLA/CaO and PLLA/MgO composites pyrolysis: Simulation analysis

Thermal degradation kinetics of the composites was studied by several analytical

approaches [29-31]. In Figure 6, the random degradation analysis plots of log[-log{1-(1-

w)1/2}] vs 1/T for experimental data of PLLA/CaO (5 wt %) (5 K min-1) and model

reactions are illustrated. It was observed that the degradation of PLLA/CaO (5 wt %)

proceeds by an nth order weight loss process at the beginning period (1/T > 0.002 K-1, T

< 220 °C), and is followed by a random degradation process (L = 2-3, where L is the

least number of repeating units of oligomer not volatilized) (1/T = 0.0020-0.00187 K-1,

T = 220-260 °C). This process is similar to that of purified PLLA [24]. However, it

should be noted that the random degradation curve gradually shifted out of the

simulation lines for random degradation (L = 2-3) toward an end stage, in which the Ea

value also approached that of PLLA-Ca pyrolysis as abovementioned. Thus, it was

considered that the degradation of PLLA/CaO (5 wt %) started with a similar

degradation behavior to that of purified PLLA, and then changed to approximate that of

PLLA-Ca with a diffusion of CaO into the PLLA matrix and an increase in temperature.

[Figure 6 goes here]

In Figure 7, the random degradation and integral analyses results of experimental

data of PLLA/MgO (5 wt %) (5 K min-1) were plotted along with model reaction

simulations. The pyrolysis of PLLA/MgO (5 wt %) composite also started according to

an nth order weight loss process (Figure 7a). Then, the degradation curve deviated from

the nth-order simulation by gradually accelerating as shown in the integral analysis plot

(Figure 7b). Considering the dominant L,L-lactide formation and the Ea value remained

at 120-130 kJ mol-1 over the whole pyrolysis, it can be assumed that a simple

16

degradation process of PLLA/MgO (5 wt %) underwent a gradual acceleration with the

diffusion of MgO into the PLLA matrix.

These kinetic results indicate that both the inorganic materials and the temperature

mutually interacted in the pyrolysis of PLLA composites in controlling the mechanism.

[Figure 7 goes here]

3.6 Mechanism of PLLA/CaO and PLLA/MgO composites pyrolysis

When PLLA/CaO (5 wt %) composite was heated in N2 flow, the decomposition of

PLLA chains will be started by reactions with CaO and followed by a backbiting

reaction from chain-end anions as shown in Scheme 1. This may occur around the CaO

particles in a lower temperature range. This reaction may correspond to the initial nth-

order weight loss process in Figure 6. Then, with increase in temperature, Ca species

will diffuse into the PLLA matrix and cause extensive random degradation and

subsequent backbiting reactions. The lower temperature may restrict the distillation of

cyclic dimers, L,L- and meso-lactide. Further, it is considered that with increase in

temperature, the unzipping depolymerization from alkoxy anion dominates so as to

produce L,L-lactide preferentially.

On the other hand, the weight loss of PLLA/MgO (5 wt %) composite started at

210 °C and continued in earnest over about 250 °C, which is 30 °C higher than

PLLA/CaO (5 wt %). At this temperature, the unzipping depolymerization from alkoxy

anion would be dominant and PLLA/MgO (5 wt %) would degrade smoothly through

the depolymerization, resulting in a high yield of L,L-lactide.

[Scheme 1 goes here]

17

It is considered that these differences between PLLA/CaO (5 wt %) and /MgO (5

wt %) composites are due to the electronegativity of Ca and Mg, which are 1.00 and

1.31 eV, respectively [32]. Since Mg is of a lower basicity than Ca, the reaction

between PLLA and MgO may merely occur at a lower level and the degradation may

also slightly proceed at temperatures less than 250 °C, at which temperature CaO can

form enough salt end structures to degrade the PLLA matrix. When the temperature

becomes higher than 250 °C, the reaction between MgO and PLLA will be accelerated,

and at the same time the unzipping depolymerization will be allowed to start, resulting

in the rapid decomposition of PLLA/MgO (5 wt %) composite and the selective L,L-

lactide formation.

Therefore, the thermal degradation of PLLA/MgO (5 wt %) composite can

regenerate the L,L-lactide predominantly at temperatures lower than 270 °C, even when

undergoing a heating process from ambient temperature. The meso-lactide formation

experienced in the heating process of PLLA/CaO (5 wt %) composite can be avoided by

using MgO instead of CaO.

3.7 Isothermal degradation of PLLA/CaO and PLLA/MgO composites

To confirm that the specific thermal degradation behavior is dependent on the

temperature, the isothermal degradation of PLLA/CaO (5 wt %) and /MgO (5 wt %)

composites was conducted in TG and Py-GC/MS. The same heating program: rapid

heating from 60 to 250 °C at a heating rate of 20 °C min-1 and holding at 250 °C for 10

min, was applied to each. Both TG curves of the composites showed almost complete

decomposition of PLLA under these conditions. Py-GC/MS profiles of pyrolyzates from

the composites are illustrated in Figure 8. The isothermal degradation of PLLA/CaO (5

18

wt %) at 250°C resulted in the production of 12.1 and 1.7 % of meso-lactide and cyclic

oligomers, respectively (Figure 8a), while the degradation of PLLA/MgO (5 wt %)

showed the production of 2.6 and 1.3 % of meso-lactide and cyclic oligomers,

respectively (Figure 8b). In these cases, the formation of D,D-lactide must be less than

the meso-lactide formation. These results were determined through the GC analysis of

the pyrolyzates on the isothermal degradation of PLLA/MgO (5wt%) at 220°C for 2h in

the glass tube oven. Figure 9 shows that almost the same result was obtained with the

pyrolyzates composed of dominant L,L-lactide, a small amount of meso-lactide, and a

trace amount of D,D-lactide. This is in accordance with our previous results in relation to

the thermal degradation of the calcium salt end-capped PLLA-Ca [24,25]. Thus, the

total yield of lactides in the isothermal degradation of PLLA/CaO (5 wt %) and /MgO

(5 wt %) should be more than about 98 % and the selectivity of L,L-lactide formation for

these two composites should be more than about 80 and 95 %, respectively.

[Figure 8 goes here]

[Figure 9 goes here]

4. Conclusions

In conclusion, the thermal degradation of PLLA/CaO (5 wt %) and /MgO (5 wt %)

composites was studied, and compared with that of pure PLLA. Though the pure PLLA

degraded with extensive racemization and production of oligomers other than L,L-

lactide, the composites showed different thermal degradation behavior without the

oligomers being produced and with limited racemization. CaO markedly decreased the

thermal degradation temperature of the composite, but resulted in some racemization at

lower temperatures of less than 250 °C. This is because of the formation of the calcium

19

carboxylate end structure and the subsequent SN2 attack from the carboxylate anion to

the asymmetrical carbon atom in the penultimate unit. In the case of PLLA/MgO (5

wt %) composite, the reaction between MgO and PLLA may be limited at low

temperatures of less than 250 °C, because of the lower basicity of Mg compared to Ca,

resulting in the racemization experienced in PLLA/CaO pyrolysis being depressed.

However, at temperatures higher than 250 °C, the reaction between PLLA and MgO

occurred smoothly causing the unzipping depolymerization, and resulting in selective

L.L-lactide production.

Therefore, a practical method for the feedstock recycling to L,L-lactide from PLLA

waste will be possible by using the PLLA/MgO composite.

Acknowledgement

This study was financially supported by the Special Coordination Funds of the Ministry

of Education, Culture, Sports, Science and Technology, and the Japanese Government.

Reference

[1] Ikada Y, Tsuji H. Macromol Rapid Commun 2000; 21(3): 117-32.

[2] Amass W, Amass A, Tighe B. Polym Int 1998; 47(2): 89-144.

[3] Anderson JM, Shive MS. Advan Drug Delivery Rev. 1997; 28 (1): 5-24.

[4] Ajioka M, Enomoto K, Suzuki K, Yamaguchi A. J Environ Polym Degrad 1995;

3(4): 225-34.

[5] Lunt J. Polym Degrad Stab 1998; 59: 145-52.

[6] Pranamuda H, Tokiwa Y, Tanaka H. Appl Environ Microbiol 1997; 63: 1637-40.

[7] Pranamuda H, Tokiwa Y. Biotech Lett 1999; 21: 901-5.

20

[8] McNeill IC, Leiper HA. Polym Degrad Stab 1985; 11: 267-85.

[9] McNeill IC, Leiper HA. Polym Degrad Stab 1985; 11: 309-26.

[10] Jamshidi K, Hyon SH, Ikada Y. Polymer 1988; 29: 2229-34.

[10] Kopinke, FD, Mackenzie K. J Anal Appl Pyrolysis 1997; 40-41: 43-53.

[12] Kopinke FD, Remmler M, Mackenzie K, Moder M, Wachsen O. Polym Degrad

Stab 1996; 53: 329-42.

[13] Aoyagi Y, Yamashita K, Doi Y. Polym Degrad Stab 2002; 76: 53-9.

[14] Babanalbandi A, Hill DJT, Hunter DS, Kettle L. Polym Int 1999; 48: 980-4.

[15] Lee SH, Kim SH, Han YK, Kim YH. J Polym Sci Part A Polym Chem 2001; 39:

973-85.

[16] Khabbaz F, Karlsson S, Albertsson AC. J Appl Polym Sci 2000; 78: 2369-78.

[17] Westphal C, Perrot C, Karlsson S. Polym Degrad Stab 2001; 73: 281-7.

[18] Brochu S, Prud’homme RE, Barakat I, Jerome R. Macromolecules 1995; 28: 5230-

39.

[19] Tsuji H, Ikada Y. Macromolecules 1992; 25: 5719-23.

[20] Radano CP, Baker GL, Smith MR. J Am Chem Soc 2000; 122: 1552-3.

[21] Noda M, Okuyama H. Chem Pharm Bull 1999; 47: 467-71.

[22] Cam D, Marucci M. Polymer 1997; 38: 1879-84.

[23] Nishida H, Mori T, Hoshihara S, Fan Y, Shirai Y, and Endo T, Polym Degrad Stab

2003; 81:515-23

[24] Fan Y, Nishida H, Hoshihara S, Shirai Y, Tokiwa Y, Endo T, Polym Degrad Stab

2003; 79: 547-62.

[25] Fan Y, Nishida H, Shirai Y, Endo T, Polym Degrad Stab 2003; 80: 503-11.

[26] Ozawa T. Bull Chem Soc Japan 1965; 38:1881-6.

21

[27] Nishida H, Yamashita M, Hattori N, Endo T, Tokiwa Y. Polym Degrad Stab 2000;

70: 485-96.

[28] Doyle CD. J Appl Polym Sci 1961; 5: 285-92.

[29] Nishida H, Yamashita M, Endo T. Polym Degrad Stab 2002; 78: 129-35.

[30] Ichihara S, Nakagawa H, Tsukazawa Y. Kobunshi Ronbunshu 1994; 51: 459-65.

[31] Simha R, Wall LA. J Phys Chem 1952; 56: 707-15.

[32] Henderson P, editor. Inorganic Geochemistry. Oxford: Pergamon Press, 1982. pp.

315.

22

Figure legends

Scheme 1. Expected degradation mechanism of PLLA/CaO and /MgO composites.

Figure 1. TG (upper) and DTA (lower) profiles on pyrolysis of PLLA, PLLA/CaO (5

wt %), and PLLA/MgO (5 wt %) composites.

23

Figure 2. Py-GC/MS chromatograms of pyrolyzates from PLLA/CaO (5 wt %) and

/MgO (5 wt %) composites during a heating process at 10 °C min-1 from 60 to 400 °C.

Figure 3. Racemization on pyrolysis of PLLA/CaO (5 wt %) (upper) and PLLA/MgO

(5 wt %) (lower) composites in different temperature ranges. Bar: amount of meso-

lactide formed (integration value of TIC peaks); line: relative content ratio (%) of meso-

lactide to total lactides.

24

Figure 4. Changes in Ea value during thermal degradation of PLLA, PLLA/CaO (5

wt %), and PLLA/MgO (5 wt %) composites.

Figure 5. Relations between Ea value and the total of meso-lactide and cyclic oligomers

contents on pyrolysis temperature of PLLA/CaO (5 wt %), PLLA/CaO (1 wt %),

PLLA/MgO (5 wt %), PLLA, and PLLA-Ca.

25

Figure 6. Random degradation simulation for pyrolysis of PLLA/CaO (5 wt %)

composite. Heating rate: 5 K min-1.

Figure 7. Random degradation (upper) and integral (lower) simulations for pyrolysis of

PLLA/MgO (5 wt %) composite. Heating rate: 5 K min-1.

26

Figure 8. Py-GC/MS chromatograms of pyrolyzates from PLLA/CaO (5 wt %) and

/MgO (5 wt %) composites on a heating process: dynamic heating from 60 to 250 °C at

heating rate of 20 °C min-1 and isothermal heating at 250 °C for 10 min.

Figure 9. GC chromatogram for volatile pyrolyzates of PLLA/MgO (5 wt %) at 220 °C

for 2 h in a glass tube oven.

2. Experimental2.1 Materials2.2 Measurements

3. Results and discussion3.2 Py-GC/MS analysis of pyrolyzates3.3 Racemization of PLLA in pyrolysis of PLLA/CaO and PLLA/MgO composites3.5 Kinetics of PLLA/CaO and PLLA/MgO composites pyrolysis: Simulation analysis3.6 Mechanism of PLLA/CaO and PLLA/MgO composites pyrolysis

Reference